venus

Confused about cuda version, cuda 10.2 or cuda 11.6 are supported or not by lightseq inference

2

when I run inference coda, cuda version cannot > cuda 11, it only works at cuda 10.2 (one issue also about this problem) eg: run code in lightseq/examples/inference/python/test/ when I...

启动后安装插件，去掉路径检测

4

  # os.makedirs( # getattr(shared.cmd_opts, 'deepdanbooru_projects_path', default_ddp_path), # exist_ok=True # ) 这里检测会出错，shared.cmd_opts.deepdanbooru_projects_path 其实是存在的，需要去掉这个代码检测报错

torch 2.3.1+cuda12.1 with flux lora train, but got the train error when use single GPU, pls help

1

## Error When I use 2 GPU to train flux lora, everything is fine, successful training~, but when I use one GPU or start with 2GPU, but use one, it...

torch 2.3.1+cuda12.1 with the latest code, but got the train error when use single GPU, pls help

1

## error When I use 2 GPU to train flux lora, everything is fine, successful training~, but when I use one GPU or start with 2GPU, but use one, it...

Firstly, thanks for your effort~ and when I use Vocos.from_pretrained("charactr/vocos-encodec-24khz"),I found it will Request huggingface everytime even I already download it, could this part code update from only download to...

`try: import torch._inductor.codecache torch._inductor.codecache.write_atomic = patched_write_atomic except: pass` I found ComfyUI-CogVideoXWrapper and ComfyUI-MochiWrapper both have apply_patched_write_atomic, which will make the main.py process bacome a lot: before:  after: I put...